The Windows Phone team announced last night that it has issued a major upgrade to its Translation mobile app that now lets users point their phones at text written in one language and have it translated into another, performs audio translation and offers visual text functionality even without a network connection. User reaction at launch has been positive, with the exception of users unsatisfied with the selection of available languages.
Language translators emphasize that this kind of technology is still far from as accurate or flexible as what a trained human mind can do and is best suited for simple task-based communication. “We have a long way to go before we will have the capabilities to build our own version of Star Trek’s ‘Universal Translator,'” wrote translator Charlie Jolly about the state of mobile translation two years ago. Despite years of technological advancement, that vision of the global mobile future remains distant still. Even heavy-duty desktop software for language translation still faces challenges.
Lightweight mobile translation is something users are interested in, though. The inclusion of offline functionality in the Windows offering is key as international data costs remain a painful problem for travelers. The addition of this functionality to Windows Phone, if it proves effective in everyday use, could put Windows in the lead in this category for now at least, compared to Apple’s minimal language translation to date and Google’s impressive offerings that require the user be online. Google’s breadth of language options will be hard to beat. Apple isn’t standing still, either. Karen Haslam wrote earlier this month at Macworld UK that Apple has applied for a patent to use data gleaned from Siri input to power language translation on iOS devices.
Augmented reality visual translation as Windows Phone now offers is something that mobile device makers have been working on for years, but translation quality, network speed and user experience have rendered their efforts ineffectual. Gathering sufficient data to learn from has also been a challenge.
